Democratizing Community Design: Generating Low-Density Expansion Plans Through Deep Learning

This research addresses the housing crisis and the need for sustainable design by using deep learning algorithms to generate optimal floor plans for expanding low-density communities in the UK. The study aims to democratize the design process, enabling residents to participate in remodeling and extending their homes, creating efficient and sustainable communities. The methodology integrates map recognition and comprehensive analysis to propose expansion plans considering spatial factors, sunlight, ventilation, and price. A neural network is trained using a dataset of 600-floor plans with privacy gradient zoning layouts. The network uses the Pix2Pix algorithm, a conditional generative adversarial network (GAN), to generate residential building floor plans (RBFPs) that conform to these privacy gradients. The research highlights the potential of AI in addressing housing challenges, providing residents with more options for home expansion and remodeling while maintaining personalized and adaptable living environments.